ORIGINAL ARTICLE Open Access

Mesocotyl Elongation is Essential forSeedling Emergence Under Deep-SeedingCondition in RiceHyun-Sook Lee1,2, Kazuhiro Sasaki2,3, Ju-Won Kang1, Tadashi Sato2,4, Won-Yong Song5

and Sang-Nag Ahn1*

Abstract

Background: Direct-seeding cultivation by deep-seeding of seeds (drill seeding) is becoming popular due to thescarcity of land and labor. However, poor emergence and inadequate seedling establishment can lead to yield lossin direct-seeding cultivation by deep-sowing. In rice, mesocotyl and coleoptile are primarily responsible for seedlingemergence from deeper levels of soil.

Results: Quantitative trait loci (QTLs) for mesocotyl and coleoptile length at 5-cm seeding depth were detectedusing 98 backcross inbred lines from a cross between Kasalath and Nipponbare. Three QTLs qMel-1, qMel-3,and qMel-6 for mesocotyl length were identified on chromosomes 1, 3, and 6, respectively, in two independent replicates.At two QTLs, qMel-1 and qMel-3, the Kasalath alleles increased mesocotyl length, whereas Nipponbare allele increased atqMel-6. The Nipponbare alleles at two QTLs (qCol-3 and qCol-5) increased the coleoptile length. Further, seeds of 54chromosome segment substitution lines (CSSLs) from the cross between Kasalath and Nipponbare sown at 5 cm soildepth showed a significant positive correlation between seedling emergence and mesocotyl elongation (r > 0.6, P < 0.0001), but not with coleoptile elongation (r = 0.05, P = 0.7). Seedling emergence of Nipponbare, Kasalath, and the 3 ofthe 54 CSSLs rapidly decreased with increasing sowing depth. Seedling emergence at seeding depths of 7 and 10 cmwas faster in Kasalath and CSSL-5 that harbored the Kasalath alleles across the qMel-1 and qMel-3 regions than in theother two CSSLs that contained a single QTL and Nipponbare alleles. CSSL-5 showed the longest mesocotyl amongthe 3 CSSLs, but no difference in coleoptile length was observed among the 3 CSSLs at seeding depths of 7 and10 cm.

Conclusion: Variation of mesocotyl elongation was found to be associated with seedling emergence at the seedingdepth of 5 cm. To our knowledge, this is the first study performed using CSSLs to detect QTLs for mesocotylor coleoptile elongation and to determine the effect of mesocotyl elongation on seedling emergence in rice.Our findings provides a foundation for developing rice cultivars that show higher seedling emergence afterdirect seeding by introgressing QTLs for mesocotyl elongation in rice breeding.

Keywords: Rice, Mesocotyl, Coleoptile, Seedling emergence, QTL, CSSLs (chromosome segment substitution lines)

* Correspondence: ahnsn@cnu.ac.kr1College of Agriculture and Life Sciences, Chungnam National University,Daejeon 305-764, South KoreaFull list of author information is available at the end of the article

© The Author(s). 2017 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made.

Lee et al. Rice (2017) 10:32 DOI 10.1186/s12284-017-0173-2

BackgroundIn Asia, two rice planting methods are used: trans-planting and direct seeding. In direct seeding, seedsare sown directly in wet/puddled soil, unpuddledsoil, or standing water (Kumar and Ladha 2011). InAsia, approximately 21% of the total rice area isused for direct seeding, and this is expected to in-crease owing to the scarcity of land, water, and labor(Pandey and Velasco 2005). Rice varieties suitablefor direct seeding should include high germinationability, seedling vigor, fast root growth, early tilleringability, and lodging resistance (Lee et al. 2002; Tang2002). Faster and uniform germination and seedlingemergence resulted in more vigorous seedlinggrowth and increasing yield in direct-seeding (Farooqet al. 2006). In direct seeding, deep seeding (drillseeding) is known to reduce damages from wildlifeand improve lodging tolerance. However, when seedsare sown deep, the seedlings need to elongate theirorgans to ensure that the plumule reaches the soilsurface. Seedling emergence is an important criterionfor determining the actual yield during direct seed-ing cultivation of rice. The coleoptile (the protectivesheath that covers the emerging shoot) and mesoco-tyl (the structure between the scutellar and coleopti-lar nodes in an embryo) of rice are primarilyresponsible for the emergence of seedlings from dee-per soil layers (Turner et al. 1982; Dilday et al.1990). Seeding depth is also an important factor fordry direct-seeding rice cultivation. Seedling growthfor temperate japonica cultivars was adversely af-fected when seeds were sown deeper than theoptimum seeding depth of 2–3 cm (Lee et al. 2002).However, shallow seeding can increase the incidencesof damage by birds, and plants might suffer lodgingafter the heading stage. Thus, obtaining informationon the optimum seeding depth for direct seeding fordiverse germplasm is necessary (Lee et al. 2002).Previous studies reported that mesocotyl and cole-

optile elongation are governed by many genetic andenvironmental factors (Takahasi 1978; Takahasi1984). Their elongation shows significant variationacross genotypes. Mesocotyl elongation is greater inindica rice than in japonica rice (Suge 1972; Taka-hashi 1978; Lee et al. 2012), whereas the coleoptileis longer in japonica cultivars than in indica culti-vars under submerged condition (Takahashi 1978).Among japonica cultivars, upland rice generally pro-duced shorter coleoptiles and longer mesocotylsthan lowland type (Chang and Vergara 1975).Quantitative trait loci (QTLs) associated with mesocotyl

and coleoptile elongation in rice have been reported. QTLsfor mesocotyl elongation were detected using various segre-gating populations from interspecific or intrasubspecific

crosses (Katsuta-Seki et al. 1996; Cai and Morishima 2002;Cao et al. 2002; Huang et al. 2010; Lee et al. 2012). Katsuta-Seki et al. (1996) reported 3 QTLs for mesocotyl elongationon chromosomes 3, 6, and 11 by using F3 populations fromthe cross between 2 indica cultivars grown in glass tubes.Cai and Morishima (2002) detected 11 QTLs for mesocotyllength by using 125 recombinant inbred lines (RILs) derivedfrom a cross between an indica cultivar and a strain of wildrice. Further, 5 QTLs for mesocotyl length were detectedus-ing an RIL population from the cross between 2 japonicacultivars by using the filter paper method (Huang et al.2010). Cao et al. (2002) detected 8 QTLs on chromosomes1, 3, 6, 7, 8, and 12 under moderate and low temperate con-ditions in a doubled haploid population derived from a crossbetween a japonica cultivar and an indica cultivar. In a pre-vious study, we detected 5 QTLs for mesocotyl length onchromosomes 1, 3, 7, 9, and 12 by using agar media and thesame backcross inbred line (BIL) populations used in thisstudy (Lee et al. 2012).QTLs for improving seedling vigor or seedling estab-

lishment, including traits for seedling emergence, seedgermination, or shoot length, and coleoptile length havebeen detected (Redoña and Mackill 1996; Zhou et al.2007; Xie et al. 2014). Redoña and Mackill (1996)detected 2 QTLs for coleoptile length and 5 QTLs formesocotyl length by using an F2:F3 population derivedfrom a cross between japonica cultivar Labelle andindica cultivar Black Gora. Further, 5 QTLs for seedvigor traits such as coleoptile length and radicle lengthand 3 QTLs for seed germination under low and normaltemperature conditions were identified using popula-tions derived from a cross between two indica cultivars(Xie et al. 2014). In addition, 9 QTLs for seedling vigortraits, including coleoptile emergence, were detectedunder two field conditions—drained and flooded soi-l—and 2 QTLs for coleoptile emergence were identifiedon chromosomes 1 and 3 under flooded soil condition(Zhou et al. 2007).In this study we aimed to show that mesocotyl elong-

ation is essential to seedling emergence under deep-seeding using chromosome segment substitution lines(CSSLs). Although several studies identified QTLs formesocotyl or coleoptile elongation, the locations andeffects of the detected QTLs varied according to thetest methods (Redoña and Mackill 1996; Zhou et al.2007; Huang et al. 2010; Xie et al. 2014). Furthermore,no study tried to characterize and detect QTLs formesocotyl under deep-seeding condition in rice. Also,study to analyze the association between mesocotylelongation and seedling emergence under various soildepth conditions is rare in crop including rice (Chung2010; Alibu et al. 2011; Lu et al. 2016).This study attempted to (1) identify QTLs associated

with mesocotyl and coleoptile elongation under the

Lee et al. Rice (2017) 10:32 Page 2 of 11

deep-seeding soil condition, and (2) to evaluate the ef-fect of mesocotyl elongation on seedling emergenceusing deep-seeded CSSLs.

ResultsPhenotypic variation of mesocotyl and coleoptileelongationNipponbare and Kasalath were selected to determine theseeding depth for detecting QTLs controlling mesocotyland coleoptile lengths. At first, Nipponbare and Kasalathseeds were placed at 5 cm and 7 cm sowing depthsbased on previous reports (Chung 2010, Alibu et al.2011). Although Nipponbare and Kasalath seeds com-pletely germinated, some seedlings showed poor growthat 7 cm depth (Additional file 1: Figure S1). Two soildepth conditions, 3 cm (shallower soil depth than 5 cm)and 10 cm (deeper soil depth than 7 cm) were added tomeasure mesocotyl and coleoptile length at 14 days aftersowing (Fig. 1). Kasalath showed exceptionally longmesocotyls than those in Nipponbare at all soil depths.Such difference in mesocotyl length between these twovarieties has also been observed using agar media (Leeet al. 2012). In contrast, the coleoptile length was signifi-cantly longer in Nipponbare than in Kasalath at all soildepths (P < 0.05). Both mesocotyl and coleoptile lengthsincreased with sowing depth in both the cultivars, butthe difference was not significant between 7 and 10 cmsowing depths. The mesocotyl lengths were 24.9 mm(seeding depth, 5 cm), 33.7 mm (7 cm), and 36.5 mm(10 cm) in Kasalath, whereas the coleoptile lengths were40.8 mm (5 cm), 44.2 mm (7 cm), and 43.2 mm (10 cm)in Nipponbare. These results suggested that the mesoco-tyl of Kasalath maximally elongates to about 35 mm,and the coleoptile in Nipponbare maximally elongates toabout 43 mm, under the experimental conditions usedin this study. Therefore, 7 cm seeding depth was consid-ered suitable for QTL analysis of mesocotyl and

coleoptile lengths in the BILs. Nevertheless, 5 cm seed-ing depth was selected for QTL analysis. Because, someseed grown insufficiently after germination hindered theaccurate phenotypic evaluation of the traits whenplanted in 7 cm and 10 cm soil depths (Additionalfile 1: Figure S1). Mesocotyl and coleoptile length ofNipponbare and Kasalath seedling were measured ex-cept insufficiently grown seedling in 7 cm and 10 cmsoil depth (Fig. 3c and d).Fifty-seven accessions from a rice diversity research set

of germplasm (RDRS) were evaluated for mesocotyl andcoleoptile length to know the association between twotraits and also to clarify whether 5 cm seeding depth wasa proper condition for analyzing phenotypic variation.This collection carries 91% of the alleles identified in the332 original rice varieties and covers most of the rangeof variation in several agro-morphological traits fromoriginal varieties (Kojima et al. 2005). Phenotypic vari-ation in mesocotyl and coleoptile lengths in 57 RDRSaccessions is shown in Fig. 2 and Additional file 2:Table S1. The rice accessions showed large variationsfor both mesocotyl and coleoptile lengths. The meso-cotyl length of the 57 accessions ranged from 0.3 mmto 37 mm, whereas the coleoptile length ranged from17.3 mm to 49.2 mm (Fig. 2a). The mesocotyl lengthshowed a significant negative correlation with the cole-optile length (r = −0.73, P < 0.001). Moreover, com-parison of mesocotyl and coleoptile elongation amongthe four groups—japonica, tropical japonica, indica-I,and indica-II, classified by Kojima et al. (2005) (Uga etal. 2009; http://www.gene.affrc.go.jp/databases-core_col-lections_wr.php)—showed that the difference amongthe groups was highly significant for both the traits(P < 0.001; Fig. 2b). The indica-I group (includingKasalath) and japonica group (including Nipponbare)showed the longest mesocotyl (27.2 mm) and coleop-tile (41.9 mm), respectively (Fig 2b). Among the

Fig. 1 Phenotypic variation of mesocotyl and coleoptile elongation at varying soil depths. a Mesocotyl and b coleoptile length of Nipponbare and Kasalathunder different burial depth in soil; 12 Seeds of Nipponbare and Kasalath were sown at 3 cm, 5 cm, 7 cm, and 10 cm soil depth and incubated at alternatetemperatures of 30 °C and 26 °C (14 h/10 h). At 14 days after sowing, the seedling were excavated and length were measured. Bars represent mean oflength with SD (n = 3). Comparisons of the varieties were made with the ANOVA test. * P < 0.05, *** P < 0.001

Lee et al. Rice (2017) 10:32 Page 3 of 11

accessions, Nipponbare and Kasalath displayed hugedifference in mesocotyl and coleoptile length indicatingthat the BILs from a cross between Nipponbare andKasalath were suitable for QTL analysis. Also, the BILswere used for mapping QTLs for a wide variety oftraits (Lin et al. 1998; Murai et al. 2002; Yamaya et al.2002). Therefore, 5 cm seeding depth was used toanalyze the genetic variation in mesocotyl and coleop-tile elongation among the 98 BILs from a cross be-tween Kasalath and Nipponbare.Two independent measurements of mesocotyl and

coleoptile lengths of the 98 BILs were obtained at the5 cm seeding depth (Fig. 3). The mesocotyl and coleop-tile lengths were not significantly different between thetwo independent experiments (r = 0.79 for mesocotyl,r = 0.66 for coleoptile; P < 0.001). The mesocotyl lengthsof Nipponbare and Kasalath were 3.3 mm and 25.1 mm,respectively, in one replicate (Fig. 3a). In the 98 BILs,the mesocotyl length ranged from 0.7 mm to 19.3 mm,with an average of 6.5 mm. The coleoptile lengths ofNipponbare and Kasalath were 33.5 mm and 28.8 mm,respectively, in one replicate (Fig. 3b.), and the averagecoleoptile length was 37.1 mm and ranged from27.6 mm to 47.2 mm in the BILs. Mesocotyl lengthshowed a highly significant negative correlation with co-leoptiles length in both the replicates (r = −0.37,P < 0.001 in 1st replicate and r = −0.20, P < 0.05 in 2ndreplicate). In BILs, no single line had longer mesocotylthan Kasalath whereas some transgressive lines showinglonger coleoptile length than Nipponbare were observed.Lack of transgressive lines for mesocotyl length is pos-sibly due to the finding that Kasalath has major mesoco-tyl increasing QTLs, qMel-1 and qMel-3 whereasNipponbare has a minor QTL whose effect was

Fig. 2 Mesocotyl and coleoptile length for 57 accessions of RDRS (a), Comparison of mesocotyl and coleoptile length in 4 rice variety groups (b) Japonica(Ja), Tropical Japonica (TJa), Indica-I (In-I), and Indica-II (In-II); Seeds were sown at 5 cm soil depth and incubated at alternate temperatures of 30 °C and 26 °C (14 h/10 h). At 14 days after sowing, the seedling were excavated and length were measured. b Means with different letters in thesame row indicate significant differences according to the Duncan multiple range test (P < 0.001). Bars represent mean of length with SE

Fig. 3 Frequency distribution of the mesocotyl length (a) andcoleoptile length (b) in 98 backcross inbred lines (BILs) derived fromNipponbare/Kasalath at 5 cm soil depth; Arrowheads indicate meanvalues for Nipponbare and Kasalath. 12 seeds of 98 BILs, Nipponbareand Kasalath were sown at 5 cm soil depth and incubated at alternatetemperatures of 30 °C and 26 °C (14 h/10 h). At 14 days after sowing,length were measured

Lee et al. Rice (2017) 10:32 Page 4 of 11

negligible in BILs. For coleoptile length, it is possiblethat Kasalath might possess a few QTL with positive ef-fects on coleoptile elongation although two minor QTLswere detected with increasing effects from Nipponbaredue to small population size of BIL, and gene interac-tions. This result is consistent with that of the previousstudies (Redoña and Mackill 1996; Xie et al. 2014).

QTL analysis for deep-seeding tolerance-related traitsQTL was detected based on logarithm of odds ratio(LOD) thresholds after the permutation test. In all, 5QTLs, including 3 for mesocotyl length and 2 for cole-optile length, were detected at the 5 cm seeding depth(Table 1, Additional file 3: Figure S2). The three QTLs(qMel-1, qMel-3, and qMel-6) for mesocotyl length weremapped near the markers—R2414, R1618, and R2123 onchromosomes 1, 3, and 6, respectively. The qMel-3 QTLwas detected in both the replicates (Rep1 and 2), andthe Kasalath alleles at two QTL loci (qMel-1 and qMel-3) contributed to an increase in mesocotyl length. TheQTL qMel-1 accounted for 13.89% and qMel-3 for37.56% and 26.97% of phenotypic variance. These QTLswere detected even when the varieties were cultured inagar (Lee et al. 2012). Two QTLs, qCol-3 and qCol-5 forcoleoptile length were mapped near the markers, C25and C597 on chromosomes 3 and 5, respectively, in onereplicate (Rep 1). In the other replicate (Rep 2), no QTLsfor coleoptile length were detected. The two QTLs,qCol-3 and qCol-5 accounted for 11.8 and 12.0% ofphenotypic variance, respectively. Nipponbare alleles atthe two QTL loci contributed to an increase in the cole-optile length. Interestingly, qMel-3 and qCol-3 showedloose linkage suggesting a possibility of combining twoQTLs via marker-assisted selection.

Effect of mesocotyl elongation on seedling emergenceIn all, 54 CSSLs were used for evaluating the effect ofQTLs for mesocotyl or coleoptile length on seedlingemergence at the 5 cm seeding depth (Fig. 4). Each

CSSL carries a single or a few Kasalath segments inthe near-isogenic background of Nipponbare. Seed-lings started to emerge at 6 days after sowing, whenthe analysis was conducted.A positive correlation was found between mesocotyl

length and seedling emergence at 6 days after sowing(r = 0.8, P < 0.0001), whereas no significant correlation wasnoted between coleoptile length and seedling emergence(P > 0.5). This result suggests that seedling emergence isdependent on mesocotyl elongation.All lines except 5 lines were not significantly different

from Nipponbare in mesocotyl and coleoptile lengths(Fig. 4). Four lines, CSSL-5, CSSL-15, CSSL33 andCSSL-54 showed significantly longer mesocotyl lengththan Nipponbare and the other CSSLs (Fig. 4b). CSSL-5showed the highest emergence percentage (70%) amongCSSLs followed by CSSL-33 (17%) and CSSL-15 (12.5%).(Fig. 4a). CSSL-5 possessed Kasalath segments acrossqMel-1 and qMel-3 and across only qMel-3 in CSSL-15.On the other hand, CSSL-33 possessed the Kasalath seg-ments across qMel-3 and qMel-7. The qMel-7 QTL wasnot detected in this study whereas it was observed in theagar condition (Lee et al. 2012). Coleoptile length of Kasa-lath and CSSL-53 was significantly shorter than Nippon-bare and the other CSSLs, but the emergence percentageof Kasalath (87.5%) was the highest (Fig. 4a and C). Thisresult suggests that seedling emergence is dependent onmesocotyl elongation.In addition, we tried to confirm the effect of

mesocotyl length on seedling emergence using 3CSSLs (Fig. 5). Lengths of mesocotyl and coleoptileand seedling emergence were investigated in 3CSSLs, possessing the Kasalath segments across theqMel-1 (CSSL-6) or qMel-3 (CSSL-15) regions andboth (CSSL-5). Seedlings of all genotypes emergedover 90% at 3 cm and 5 cm seeding depths from 8to 9 days after sowing (Fig. 5a). At 7-cm seedingdepth, Kasalath and CSSL-5 showed faster emer-gence than the other genotypes. At 10-cm seeding

Table 1 Characteristics of QTLs for mesocotyl and coleoptile length in 98 BILs

Locus a Chr. Nearest marker Position Replication b LOD score c R2 (%) d Additiveeffect

Positive allele e

qMel-1 1 R2414 39.6- Mb Rep 2 3.44 13.89 4.34 K

qMel-3 3 R1618 30.7-Mb Rep 1 12.09 37.56 6.41 K

R1618 30.7-Mb Rep 2 7.18 26.97 6.46 K

qMel-6 6 R2123 11.68-Mb Rep 2 2.81 8.28 3.89 N

qCol-3 3 C25 3.97-Mb Rep 1 3.06 11.8 2.96 N

qCol-5 5 C597 0.26-Mb Rep 1 3.20 12.0 2.70 Na QTLs were designated as “qMel chromosome number” and “qCol chromosome number”b QTLs were detected under soil conditions in two replications, respectivelyc Maximum LOD score over threshold significance level at P < 0.05d Proportion of the phenotypic variation explained by the nearest marker of QTLe K and N respectively represent the positive effects of QTL contributed by Kasalath and Nipponbare alleles

Lee et al. Rice (2017) 10:32 Page 5 of 11

depth, no seedlings of Nipponbare, CSSL-6, andCSSL-15 emerged even 14 days after sowing,whereas Kasalath and CSSL-5 seedlings emergedand showed 78 and 36% seedling emergence after14 days, respectively (Fig. 5a). The coleoptile length

of CSSL-5 were not significantly different fromthose of Nipponbare or the other two CSSLs at allseeding depth. On the contrary, the mesocotyllength of CSSL-5 showed the longest length among3 CSSLs at 5-cm, 7-cm and 10-cm seeding depth

Fig. 4 Seedling emergence percentage at 6 days after sowing (a), mesocotyl (b) and coleoptile length (c) of 54 CSSLs (SL); 12 seeds of 54 CSSLs,Nipponbare and Kasalath were sown at 5 cm soil depth and incubated and the numbers of emerged seedling from the soil surface werecounted. At 14 days after sowing, the length were measured. Each column represents emergence percentage or mean of length ± SE * P < 0.05and **P < 0.01 versus Nipponbare (Dunnett’s multiple comparison test, two replicates)

Lee et al. Rice (2017) 10:32 Page 6 of 11

(Fig. 5b). These results indicate that a significant re-lationship exists between mesocotyl elongation andseedling emergence in deep-seeded plants.Additionally, these results also show that the 2 QTLs,

qMel-1 and qMel-3 act additively in distinct or comple-mentary pathways in controlling mesocotyl elongationunder this study condition.

DiscussionIn rice, seeding depth for direct seeding is an import-ant factor for ensuring seedling establishment. Lee etal. (2002) recommended a 2–3-cm sowing depth forjaponica rice. When the seeding depth is >3 cm,seedling emergence was markedly delayed (Kawateiet al. 1963; Murai et al. 1995; Luo et al. 2007; Chung

Fig. 5 Time course of seedling emergence (a) and mesocotyl and coleoptile length (b) of 3 CSSLs, Nipponbare and Kasalath in four soil burialdepth; (a) Lines and bars represent the mean with SE of 5 lines, Nipponbare(Ni), Kasalath (Ka), CSSL-6 (SL6), CSSL-15 (SL15), and CSSL-5 (SL5).Seeds were sown at 3 cm, 5 cm, 7 cm, and 10 cm soil depth and the number of emerged seedling form soil surface were counted daily up to14 days after sowing. b At 14 days after sowing, length of mesocotyl and coleoptile were measured. Means of length ± SE (n = 3) with differentletters in the same row indicate significant differences according to the Duncan multiple range test (P < 0.05)

Lee et al. Rice (2017) 10:32 Page 7 of 11

2010; Alibu et al. 2012). Moreover, when seeds weresown deeper than 5 cm, the first leaf developed underthe soil surface, and seedling establishment decreasedremarkably (Chung 2010). Similar result was obtainedin this study: seedling emergence speed and percent-age at 7 cm and 10 cm soil depths were lower thanthose at the seeding depth of 3 cm in all the geno-types (Fig. 5). Although 3-cm seeding depth might beoptimal for direct seeding based on these results,many studies have used seeding depths of less than3 cm, which caused seed desiccation and floating andseedling lodging (Lee et al. 2002, Kumar and Ladha2011). Genotype differences for seedling emergenceand coleoptile length were more pronounced at 5 cmthan at 3 cm seeding depth (Figs. 1 and 5). The twoparents and the 3 CSSLs showed >90% seedling emer-gence within 7 days when the seeding depth was3 cm, whereas Kasalath and CSSL-5 showed >90%emergence and the other three lines reached a plateauat 80% emergence at 5 cm seeding depth (Fig. 5).These results indicate that soil depth of 3–5 cm isappropriate for direct seeding of cultivars that showgood mesocotyl elongation without affecting seedlingestablishment. Based on these findings, we selected5 cm soil depth for screening seedling emergencetraits under the deep-seeding condition.Mesocotyl and coleoptile lengths are directly related

with seedling emergence in deep seeding and enhancedmesocotyl or coleoptile elongation is associated with bet-ter seedling emergence and establishment (Turner et al.1982; Murai et al. 1995; Luo et al. 2007; Chung 2010;Alibu et al. 2012). Murai et al. (1995) reported that a re-lationship exists between seedling emergence ability andlengths of leaf, mesocotyl, coleoptile, and leaf internodeusing dwarf lines under 7 cm soil depth. Chung (2010)reported positive correlations between the rate of seed-ling emergence and mesocotyl or coleoptile length at5 cm seeding depth by using 116 Korean weedy rice ac-cessions. Alibu et al. (2012) suggested that upland riceaccessions with long mesocotyl showed better seedlingemergence than those with long coleoptile under the drydirect-seeding condition. Lu et al. (2016) showed thatlong mesocotyl rice accessions had higher emergencerate than that of the short one at 2 and 5-cm sowingdepth. In addition to the seeding depth, in dry seeding,the lengths of mesocotyl and coleoptile can be affectedby moisture content. Under submergence, coleoptilegrowth was stimulated, although no increase in mesoco-tyl length was observed (Takahashi 1978; Alibu et al.2011). This suggests that mesocotyl and coleoptileelongation is important for seedling emergence depend-ing on the environment. In our study, a positive correl-ation between mesocotyl length and seedling emergencewas observed (Figs. 4 and 5). Seedlings of Kasalath and

CSSL-5 that have long mesocotyls emerged consider-ably faster than those of the other genotypes thathave short mesocotyl at the 7 and 10 cm seedingdepths (Figs. 4 and 5). At 7 cm seeding depth, seed-lings of Kasalath and CSSL-5 emerged over 70% in 7-8 days after sowing, whereas those of the other geno-types and Nipponbare showed 70% emergence at 10–11 days after sowing. This delayed seedling emergenceof up to 3–4 days might be unfavorable for seedlingestablishment and competition with weeds.Evaluation of RDRS showed genotypic variation of

mesocotyl and coleoptile length (Fig. 2). Japonica acces-sions tend to have shorter mesocotyls than indica,whereas japonica showed the longest coleoptiles amongfour groups. In this study, indica-I accessions had thelongest mesocotyls among the four groups, followed bythe tropical japonica group (Fig. 2b). These results areconsistent with the previous findings (Suge 1972; Taka-hashi 1978, Sato 1987).QTL analysis was conducted for mesocotyl and cole-

optile length of seedlings sown at 5 cm soil depth. Meas-uring mesocotyl and coleoptile traits in the soil isdifficult mainly due to the environmental influence andexperimental errors. Several studies attempted to mapQTLs for mesocotyl and coleoptile length by using theglass tube method (Katsuta-Seki et al. 1996), slant-boardtest (Redoña and Mackill 1996), filter paper with distilledwater (Huang et al. 2010), agar medium (Lee et al.2012), and filter paper on agar medium (Xie et al. 2014).However, QTL analysis for mesocotyl or coleoptileelongation at the 5 cm seeding depth by using CSSLs inrice has not yet been performed. We detected 5 QTLsunder field condition in two replicates. Interestingly,QTLs for mesocotyl elongation were commonly mappedto chromosomes 1 and 3 for different mapping popula-tions and under different experimental conditions. TheQTL qMel-3 was located on the long arm of chromo-some 3, and this QTL interval overlapped with regionsof QTLs reported in previous studies (Katsuta-Seki et al.1996; Redoña and Mackill 1996; Cai and Morishima2002; Cao et al. 2002; Huang et al. 2010). Usinggenome-wide association study (GWAS) associationmapping, Wu et al. (2015) and Lu et al. (2016) identifiedsome candidate genes controlling mesocotyl elongationon chromosome 1 and 3. Coleoptile QTLs were detectedon chromosomes 3 and 6 by Redoña and Mackill (1996)and on chromosomes 5, 8, and 11 by Xie et al. (2014).On chromosomes 1 and 3, two QTLs for coleoptilelength were identified under flooded soil condition(Zhou et al. 2007). Interestingly, qCol-3 and qCol-5 thatwere detected in the present study did not overlap withthe coleoptile length QTLs that was reported previously;this could be attributed to the different genetic materialsand test methods used.

Lee et al. Rice (2017) 10:32 Page 8 of 11

Among 54 CSSLs, two lines, CSSL-5 and CSSL-33showed significantly longer mesocotyl than Nipponbare.CSSL-5 showed the highest emergence ratio (70%)followed by CSSL-33 (17%) (Fig. 4a and b). CSSL-5 andCSSL-33 shared qMel-3 suggesting that qMel-3 is amajor QTL. In contrast, qMel-7 is minor and affected bytesting methods based on the finding that the qMel-7QTL was detected in the agar condition (Lee et al. 2012)but not in the soil condition in this study. However,CSSL-5 seedlings were shorter in mesocotyl length andlower in emergence rate than Kasalath. The results fromthis study indicate that the qMel-3 QTL with a highadditive effect has the potential to enhance early seed-ling emergence in combination with the qMel-1, qMel-7and other QTLs not detected. Also, information on theinteraction among these QTLs will be necessary tounderstand the genetic control of mesocotyl length.These findings will contribute to our understanding ofthe genetic control of mesocotyl and coleoptile elong-ation, and seedling emergence and provides informationon QTL regions associated with early seedling emer-gence for rice breeding programs.

ConclusionsWe showed that the variation in mesocotyl elongation isassociated with seedling emergence under the deep-seeding condition. Genetic analysis using BIL and CSSLplants confirmed that two major QTLs, qMel-1 andqMel-3 for mesocotyl elongation have the potential toenhance early seedling emergence at 5 cm seedingdepth. This is the first report to show the effect of twoQTLs for mesocotyl elongation detected both in agarmedia and deep-seeding soil on seedling emergenceusing CSSLs using deep-seeding soil condition. (Lee etal. 2012). High-resolution mapping and cloning of theseQTLs should be performed to understand the geneticcontrol of mesocotyl elongation to ensure higher seed-ling emergence and to develop deep-seeding-tolerantrice cultivars for direct seeding by using marker-assistedselection.

MethodsPlant materialsFifty-seven rice accessions selected from the RDRS ofgermplasm collection were used to detect variation inmesocotyl and coleoptile elongation (Oryza sativa L.)(Kojima et al. 2005; Lee et al. 2012; Additional file 2:Table S1). QTLs for mesocotyl and coleoptile elongationwere detected and the effect of mesocotyl elongation onseedling emergence was determined by using 98 BILsand 54 CSSLs developed from the backcross of theNipponbare/Kasalath (Lin et al. 1998). Seeds of BILsand CSSLs were provided by Rice Genome Resource

Center (RGRC), Japan (http://www.rgrc.dna.affrc.go.jp/stock.html).The effect of QTLs for mesocotyl elongation qMel-1

and qMel-3 on seedling emergence was determined byselecting 3 CSSLs—CSSL-6, CSSL-15, and CSSL-5 (Lee etal. 2012). The genotype data of the CSSLs were referredfrom a database of RGRC (http://www.rgrc.dna.affrc.-go.jp/ineNKCSSL54.html). CSSL-6 and CSSL-15 car-ried the QTLs qMel-1 and qMel-3, respectively.CSSL-6 contained a 34-cM Kasalath segment flankedby restriction fragment length polymorphism (RFLP)markers C86-C112 carrying qMel-1 on chromosome1. CSSL-15 contained a 61-cM Kasalath segmentflanked by RFLP markers R19–R1925 carrying qMel-3on chromosome 3 in the Nipponbare background.CSSL-5 carried both QTLs qMel-1 and qMel-3 in theNipponbare background. In CSSL-5, 2 Kasalath seg-ments on qMel-1 and qMel-3 were introgressed, a34-cM Kasalath segment near the RFLP makersC1370–2018, including qMel-1, 43-cM Kasalath seg-ment near RFLP markers C136–R1925, in the Nip-ponbare background.RDRS, BIL, and CSSL plants were grown in the experi-

mental lowland field of the Graduate School of Life Sci-ences, Tohoku University, Kashimadai, Osaki, MiyagiPrefecture, Japan. The panicles of these plants were har-vested at 40–50 days after heading and then dried in awell-ventilated space for 3 months. Dried seeds werecollected from the panicle and placed in paper enve-lopes. All of the seeds were completely germinated in30 °C imbibitions for 4 days. For maintaining the germi-nability, the seed envelopes were packed in plastic bagssealed with silica gel and stored at 4 °C in a refriger-ator until use.

Measurement of mesocotyl and coleoptile elongationMesocotyl and coleoptile elongation was measured andseedling emergence under the soil condition was evalu-ated by sowing 12 good-quality seeds from each RDRSaccession, BILs, and CSSLs at various soil depths. Seedsof Nipponbare, Kasalath, and 3 CSSLs were buried at adepth of 3, 5, 7, and 10 cm in plastic pots (diameter,11 cm; height, 15 cm) in soil containing chemicalfertilizer (Nursery culture soil No. 3; Mitsui-Toatsu,Japan). For determining phenotype variation in mesoco-tyl and coleoptile length, seeds of RDRS, 98 BILs and 54CSSLs were buried in 5 cm soil depth. The plastic potswere kept in plant growth cabinets at alternate tempera-tures of 30 °C and 26 °C (14 h/10 h). The water level ofthe soil was maintained 2 cm from the bottom of thepots to maintain adequate soil moisture for germinationand seedling growth. At 14 days after sowing, the seed-lings were carefully excavated and washed, and theirlengths were measured. Seedlings insufficiently grown

Lee et al. Rice (2017) 10:32 Page 9 of 11

were excluded from measurement. In this study, meas-urement was made on seedlings showing 3rd-4th leavesexpansions at 14 days after sowing. Mesocotyl (the dis-tance from the basal part of the seminal root to thecoleoptilar node) and coleoptile (the distance from thecoleoptilar node to the tip of the coleoptile) lengths ofeach seedling were measured using a ruler. Seedlingsthat germinated but grew insufficiently were excludedfrom the length measurements. The numbers ofemerged seedlings from the soil surface were counteddaily up to 14 days when seedling emergence wasthought to be complete. The experiment was performedin a completely randomized design with two or threereplicates of each line.

Genotype data of BILs and QTL analysisThe QTLs controlling mesocotyl and coleoptile elongationin BILs were mapped by using genotype data generatedusing 245 RFLP markers (http://www.rgrc.dna.affrc.go.jp/ineNKBIL98.html). Linkage analysis of the genotypic datawas performed using the Kosambi function of Mapmaker/EXP 3.0 software (Lander et al. 1987). QTL analysis wasperformed using composite interval mapping (CIM) byusing the QTL Cartographer version 2.5 software (Wanget al. 2012). CIM analysis was performed using forward-backward stepwise regression model 6 with a 10-cM win-dow size. The LOD threshold significance level (P < 0.05)was determined by computing 1000 permutations. TheQTL positions were assigned to the points of maximumLOD score in the target regions. The percentage of totalphenotypic variance noted for each QTL was estimatedon the basis of the R2 value.

Additional files

Additional file 1: Figure S1. Seedling of Nipponbare and Kasalathgrowing at 5 cm and 7 cm soil depth condition; 50 seeds of Nipponbareand Kasalath were sown at 5 cm and 7 cm soil depth and incubated atalternate temperatures of 30 °C and 26 °C (14 h/10 h). At 21 days aftersowing, the seedling were excavated and the length of coleoptile (Col)and mesocotyl (Me) were measured. Arrows indicate mesocotyl.(TIFF 6962 kb)

Additional file 2: Table S1. Variation of the mesocotyl and coleoptilelength for 57 rice accessions from RDRS collection. (PDF 73 kb)

Additional file 3: Figure S2. QTL cartographer LOD peak for mesocotyland coleoptile length at 5 cm soil depth; A QTL Cartographer plotobtained following composite interval mapping (CIM) using 2 replicates(Rep1 and 2). Significance of QTL is indicated by LOD score above thethreshold values determined by permutation analysis at a significant levelof P < 0.05. The graph below shows the additive effects for each of theQTL identified. Marker designations are given at the bottom and thegenetic distances (cM) are given above the horizontal line. (TIFF 4914 kb)

AcknowledgementsThis work was carried out with the support of and Basic Science ResearchProgram through the National Research Foundation of Korea (NRF) funded bythe Ministry of Education (2015R1A6A3A01060388) and "Cooperative Research

Program for Agriculture Science & Technology Development (Project No. PJ011048)", Rural Development Administration, Republic of Korea.

Authors’ contributionsHL conceived of the study and carried out molecular genetic analysis andstatistical analysis, and wrote manuscript. KS helped to design research andperformed statistical analysis. JK and SW advised to draft the manuscript. TSparticipated in the design of the study. SA designed research and helped todraft the manuscript. All authors read and approved the final manuscript.

Competing interestsThe authors declare that they have no competing interests.

Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.

Author details1College of Agriculture and Life Sciences, Chungnam National University,Daejeon 305-764, South Korea. 2Graduate School of Life Sciences, TohokuUniversity, 2-1-1 Katahira, Aoba-ku, Sendai 980-8577, Japan. 3Institute forSustainable Agro-ecosystem Services, Graduate School of Agricultural andLife Sciences, The University of Tokyo, 1-1-1 Midoricho, Nishitokyo, Tokyo188-0002, Japan. 4Graduate School of Agricultural Science, Tohoku University,1-1 Amamiya-machi, Tsutsumidori, Aoba-ku, Sendai, Miyagi 980-8555, Japan.5POTECH-UZH Cooperative Laboratory, Department of Integrative Bioscienceand Biotechnology, Pohang University of Science and Technology, Pohang790-784, South Korea.

Received: 23 March 2017 Accepted: 10 July 2017

ReferencesAlibu S, Saito Y, Shiwachi H, Irie K (2011) Relationship between coleoptile and

mesocotyl elongation of upland rice (Oryza sativa L.) seedlings undersubmergence and soil-sand culture. Afr J Agric Res 6(31):6463–6472

Alibu S, Saito Y, Shiwachi H, Irie K (2012) Genotypic variation in coleoptile ormesocotyl lengths of upland rice (Oryza sativa L.) and seedling emergence indeep sowing. Afr J Agric Res 7(47):6239–6248

Cai HW, Morishima H (2002) QTL clusters reflect character associations in wildand cultivated rice. Theor Appl Genet 104:1217–1228

Cao L, Zhu J, Yan Q, He L, Wei X, Cheng S (2002) Mapping QTLs withepistasis for mesocotyl length in a DH population from indica-japonicacross of rice (Oryza sativa). Chinese J Rice Sci 16(3):221–224 (in Chinesewith English abstract)

Chang T, Vergara BS (1975) Varietal diversity and morpho-agronomiccharacteristics of upland rice. In: Major research in upland rice, InternationalRice Research Institute, Philippines

Chung NJ (2010) Elongation habit of mesocotyls and coleoptiles in weedy ricewith high emergence ability in direct-seeding on dry paddy fields. CropPasture Sci 61(11):911–917

Dilday RH, Mgonja MA, Amonsilpa SA, Collins FC, Wells BR (1990) Plantheight vs. mesocotyl and coleoptile elongation in rice: Linkage orpleiotropism? Crop Sci 30:815–818

Farooq M, Barsa S, Wahid A (2006) Priming of field-sown rice seed enhancesgermination, seedling establishment, allometry and yield. Plant Growth Regul49:285–294

Huang C, Jiang S-K, Feng L-L, Xu Z-J, Chen W-F (2010) Analysis of QTLs formesocotyl length in rice (Oryza sativa L.) Acta Agron Sin 36(7):1108–1113

Katsuta-Seki M, Ebana K, Okuno K (1996) QTL analysis for mesocotyl elongation inrice. Rice Genetics Newsletter 13:126

Kawatei K, Hoshikawa K, Takashima Y (1963) On germination and subsequentearly growth of seedling in direct seeding culture of paddy rice on uplandfield. Jpn J Crop Sci 31(3):267–271

Kojima Y, Ebana K, Fukuoka S, Nagamine T, Kawase M (2005) Development of anRFLP-based rice diversity research set of germplasm. Breed Sci 55:431–440

Kumar V, Ladha JK (2011) Direct Seeding of Rice: Recent developments andfuture research needs. Adv Agron 111:297–413

Lander ES, Green P, Abrahamson J, Barlow A, Daly MJ, Lincoln SE, Newburg L(1987) MAPMAKER: An interactive computer package for constructing

Lee et al. Rice (2017) 10:32 Page 10 of 11

primary genetic linkage maps of experimental and natural populations.Genomics 1:174–181

Lee M-H, Kim JK, Kim SS, Park ST (2002) Status of dry-seeding technologies forrice in Korea. In: Pandey S, Mortimer M, Wade L, Tuong TP, Lopez K, Hardy B(eds) Direct seeding: Research strategies and opportunities. International RiceResearch Institute, Philippines, pp 161–176

Lee HS, Sasaki K, Higashitani A, Ahn SN, Sato T (2012) Mapping andcharacterization of quantitative trait loci for mesocotyl elongation in rice(Oryza sativa L.). Rice 5: 13

Lin SY, Sasaki T, Yano M (1998) Mapping quantitative trait loci controlling seeddormancy and heading date in rice, Oryza sativa L., using backcross inbredlines. Theor Appl Genet 96:997–1003

Lu Q, Zhang M, Niu X, Wang C, Xu Q, Feng Y, Wang S, Yuan X, Yu H, WangY, Wei X (2016) Uncovering novel loci for mesocotyl elongation andshoot length in indica rice through genome-wide association mapping.Planta 243:645–657

Luo J, Tang S, Hu P, Aleman L, Jiao G, Tang J (2007) Analysis on factors affectingseedling establishment in rice. Rice Sci 14(1):27–32

Murai M, Sato S, Ise K, Semihira Y (1995) Effects of the dwarfing gene from dee-geo-woo-gen and others on emergence ability under deep seedingcondition in rice. Breed Sci 45:163–171

Murai K, Lin S, Yano M, Nagamine T (2002) Mapping quantitative trait locicontrolling seed loggevity in rice (Oryza sativa L.) Theor Appl Genet104(6):981–986

Pandey S (2005) Velasco L (2005) Trends in crop establishment methods in Asiaand research issues. In: Toriyama K, Heong KL, Hardy B (eds) Rice Is Life:Scientific perspectives for the 21st Century, p.178–181. Proceedings of theworld rice research conference, Tsukuba

Redoña ED, Mackill DJ (1996) Mapping quantitative trait loci for seedling vigor inrice using RFLPs. Theor Appl Genet 92:395–402

Sato YI (1987) Character association patterns among Japonica rice varieties ofAsian origin. In: Hsieh SC (ed) Crop exploration and utilization of geneticresources. Taichung District Agricultural Improvement Station, Taiwan

Suge H (1972) Mesocotyl elongation in japonica rice: Effect of high temperaturepre-treatment and ethylene. Plant Cell Physiol 13:401–405

Takahashi N (1978) Adaptive importance of mesocotyl and coleoptile growth inrice under different moisture regimes. Aust J Plant Physiol 5:511–517

Takahashi N (1984) Seed germination and seedling growth. In: Tsunoda S,Takahashi N (eds) Biology of rice. Japan Science Society press Elsevier,Japan, pp 71–88

Tang S (2002) Seedling broadcasting in china: and overview. In: Pandey S,Mortimer M, Wade L, Tuong TP, Lopez K, Hardy B (eds) Direct seeding:Research strategies and opportunities. International Rice Research Institute,Philippines, pp 177–184

Turner FT, Chen CC, Bollich CN (1982) Coleoptile and mesocotyl lengths insemidwarf rice seedlings. Crop Sci 22:43–46

Uga Y, Ebana K, Abe J, Morita S, Okuno K, Yano M (2009) Variation in rootmorphology and anatomy among accessions of cultivated rice (Oryza sativaL.) with different genetic backgrounds. Breed Sci 59:87–93

Wang S, Basten CJ, Zeng ZB (2012) Windows QTL Cartographer 2.5. Departmentof Statistics, North Carolina State University, Raleigh, NC. http://statgen.ncsu.edu/qtlcart/WQTLCart.htm.

Wu J, Feng F, Lian X, Teng X, Wei H, Yu H, Xie W, Yan M, Fan P, Li Y, Ma X, Liu H,Yu S, Wang G, Zhou F, Luo L, Mei H (2015) Genome-wide association study(GWAS) of mesocotyl elongation based on re-sequencing approach in rice.BMC Plant Biol 15:218

Xie L, Tan Z, Zhou Y, Xu R, Feng L, Xing Y, Qi X (2014) Identification and finemapping of quantitative trait loci for seed vigor in germination and seedlingestablishment in rice. J Integr Plant Biol 56:749–759

Yamaya T, Obara M, Nakajima H, Sasaki S, Hayakawa T, Sato T (2002) Geneticmanipulation and quantitative-trait loci mapping for nigtogen recycling inrice. J Exp Bot 53(370):917–925

Zhou L, Wang J-K, Yi Q, Wang Y-Z, Zhu Y-G, Zhang Z-H (2007) Quantitative trait locifor seedling vigor in rice under field conditions. Field Crop Res 100:294–301

Lee et al. Rice (2017) 10:32 Page 11 of 11